Information Technology Reference
In-Depth Information
firing event that includes at least 90% of the recorded neurons (Maeda
et al.
1995; Marom and Shahaf 2002; Segev
et al
. 2002; Mukai
et al
. 2003). However,
to our view, an important aspect of a SBE is not the relative number of activated
neurons, i.e. the scale on which the collective firing takes place, but the fact that
the SBE represents an initiated
activation
of a sub-network in the culture
(assumed by a localized initiation zone) followed by a
propagation
of the activity
from the initiating (leading) sub-network towards the rest of the network. These
two terms, activation and propagation, lead together to the phenomenon of
synchronization of cells when the activity cascades along the network map of
connectivity (Raichman and Ben-Jacob 2008). By our definition a SBE is any
synchronized event in the network that is a result of an activated sub-network and
not a random occurrence of nearly-timed spikes. We also do not limit the size of
SBEs to include only a large percentage of the recorded neurons, which enables
us to analyze synchronized events of just a few neurons as well.
12.5. The Characterization of the SBEs
The dynamical state of a network at a certain time window can be measured by
several key parameters, such as the rate of SBE generation, fraction of active
neurons and SBE intensity. In Raichman and Ben-Jacob (2008) we had shown
that the intensity of a SBE, either measured by the number of active neurons or
the SBE duration, is positively correlated to the time passed from the preceding
event. Thus, the intensity of the SBE reflects the network "readiness" to fire after
some time integral of silenced activity. Other two important parameters are the
activation time τ
act
, which is the time it takes to trigger the neurons (from the first
firing neuron to the initiation of the last one neuron), and the spike-train duration
time τ
st
, which is the average duration of the spike-trains of the firing neurons in
the burst. The activation time reflects the time it takes to activate the entire
network from end-to-end of the recording area, disregarding the intensity of
firing of the individual neurons, and the spike-train duration time reflects the
intensity of firing of the neurons, disregarding the temporal order in which
neurons were triggered. The exact method of calculating τ
act
and τ
st
is described
in Raichamn and Ben-Jacob (2008).
Early developing networks show dramatic changes in several key parameters,
reflecting their ongoing maturation and increase in connectivity. Developing
networks begin to exhibit spiking activity at around 3-5 DIV. Synchronized
bursting with an increasing number of neurons shortly follows the activation of
the network. In Fig. 12.4 we show the changes in the SBE frequency and
intensity in an immature developing culture, as reflected by several key
